AbYSS Algorithm Research And Its Application On Constellation Optimization Design

In our daily life, there are many MOPS(Multiobjective Optimization Problem) in fields such as engineering, science and technology, aerospace, economy and so on. In MOPs, some objectives conflict and compete each other. If you want to enabe one objeictve to be better, but other objectives became worse. For example, when we want to design a new product, we hope that the coming products have perfect function, lower product price, and better manufacturability, reliability and maintainability as well. The three objectives maybe confict, so we need a scheme which make the objectives compatible great. For explaining the MOP, we take into the investment proplem, we conside that we do my best to use the least money to obtain the best benefit in shortest period. According to what's disscuss above, we define the MOP as searching a resolution set in the some feasibility space that make the objectives compatible so that we obtain the largest benefit. With the development of industry, with the improvement of science and technology, with the continuous extension of people's cognition scope, the technology of multi-objective optimization has been developed pefectly in recent years. One of important improvements is Scatter Search algorithm. It is important and necessary to continue research Scatter Search algorithm in theory and engneering.In this paper, on the basis of the feature of AbYSS and constellation optimization design, we apply AbYSS(Archive-Based hYbrid Scatter Search) into constellation optimization design to resolve the problem of satellite constellation optimization design. First of all, we analysize the important significance of the research of Scatter Search algorithm and constellation opimization design. Then we introduce several basic method of Scatter Search algorithm, which include diverse initial population generation method, improvement method, reference set update method,subset generation method, and solution combination method. At last, we have studied the pattern of Scatter Search, multi-objective Scatter Search, physical model of constellation, optimization parameters, optimization objectives and so on. The mainly studied work as follow:1) We stuty deeply the single-objectiove Scatter Search algorithm and the mliti-objective Scatter Search algorithm. The single-objective Scatter Search has some shortcomings of initial population scattered less evenly over the feasible space, the poor ability of finding global optimal solutions and the large number of objective function evaluations, so that we design an algorithm called Orthogonal Scatter Search (OSS) algorithm by add Orthogonal design. The experimental results indicate that this algorithm can overcome the shortcomings of Scatter Search and use less number of objective function evaluations to find more accurate global optima.2)Through analyzing the optimization parameters and objectives of constellation optimization, encoding of optimization parameter and AbYSS are got and applied into constellation optimization to resove this problem. The results show that AbYSS is effective for constellation optimization problem and its optimization speed is faster than NSGA- II. Meantime, we obtain a smooth and even Pareto front.3)In order to display the results with visualization technology to readers,we have developed a simulation platform to display timely the architecture and process of constellation and the coverage of constellation from space to earth. It has realistic and guidance meaning to us.In a word, there is a deep analysis and discuss about Scatter Search algorithm and constellation optimization design in this paper. It not only proposes some effective improvement measures and extends the applied fields, but also resolves the constellation optimization design problem. In the end of this paper, we conclude what's we do and give the further research derections.